Syslinux - Jan 2014 - USB boot problems on Gigabyte GA-M55Plus-S3G

In message <52D97530.5040706 at zytor.com>, 
"H. Peter Anvin" <hpa at zytor.com> wrote:
>On 01/09/2014 03:35 PM, Ronald F. Guilmette wrote:
>> 
>> 	SYSLINUX 5.10 2013-06-04 ...
>> 
>> 	Failed to load ldlinux.c32
>> 	Boot failed: please change disks and press a key to continue.
>> 
>> In the case of the Ultimate Boot CD, the messages are a bit different:
>> 
>> 	SYSLINUX 4.07 2013-07-25 ...
>> 
>> 	ERROR: No configuration file found
>> 	No DEFAULT or UI configuration found
>> 
>> As far as I have been able to gather, this sort of thing is a common
>> problem on various Gigabyte motherboards.
>> 
>
>What are the letters that should have been printed between SYSLINUX and
>the date?  There should be "EDD" or "CHS".  This matters
significantly.

STOP THE PRESSES!

Ummmm... WOW!  I had to dig back into my e-mail outbox in order to
reestablish enough context to be able to even try to answer your
question, and it is Good that I did, because as a side-effect of
trying to answer your question, I found something TOTALLY ASTONISHING.
But we'll get to that in a second...

To answer your question...

    *) with respect to Ultimate Boot CD (5.2.7), the answer is "CHS".

    *) with respect to OpenELEC (3.2.3, I think), the answer is "CHS".

    *) with respect to Clonezilla (2.2.0) I can't easly check anymore,
       because I overwrote the stick that contained that (while performing
       some of the other tests people asked me to conduct).  If it is
       important, I will get 2.2.1 and try to answer with that, or I'll
       see if I can find an archived copy of 2.2.0.  Let me know if it is
       that important.


The REALLY ASTONISHING thing that I found while trying to answer your
question is that apparently, what Gigabyte tech support told me was
either (a) completely inaccurate or else (b) was somehow misinterpreted
by me.  What I believe they told me what that in order for *any* kind
of device to be booted by the GA-M55Plus-S3G that device *MUST* show up
in the BIOS Boot Priority list.  But as I have now seen, as a result of
trying to answer your question, is that THAT IS SIMPLY NOT TRUE!  I checked,
carefully, and even though neither of the sticks I have that contain (a)
Ultimate Boot CD (5.2.7), nor (b) OpenELEC (3.2.3) show up in the BIOS
Boot Priority list, the BIOS will, nontheless, at least TRY to at least
BEGIN the booting process off those sticks... even if the whole process
does fail later on, for reasons that are still being puzzled out.

At this point I could be... and perhaps should be... REALLY angry at
Gigabyte tech support for giving me what appears to have been dead wrong
information, but I'm not going to judge until I have time to go back and
fish out their actual e-mails to me and re-read them all, you know, just
to make sure that I didn't simply misinterpret what they told me.

In any case, other experiments that I have just now done confirm that the
real controlling factor(s), for this motherboard, are the settings of
the BIOS's "First Boot Device", "Second Boot Device",
and "Third Boot
Device" parameters.  If there is both a regular (sATA) drive and an even
potentially/theoretically bootable USB stick installed in the machine...
even one that has only one partition, and where that partition might NOT
have the bootable flag set... and if the First, Second, and Third Boot
Devices are set as follows:

	First: CDROM
	Second: USB-HDD
	Third: Hard Disk

then (in the absence of any CD/DVD) the BIOS will make an attempt to boot
from the USB stick, before trying the sATA drive, EVEN IF the USB stick in
question DOES NOT appear in the Boot Priority list.  And also, as one would
expect, if the BIOS settings are instead:

	First:	CDROM
	Second:	Hard disk
	Third:	USB-HDD

then (in the absence of any CD/DVD) then the BIOS will try to boot from the
sATA Drive first, and will do so, assuming that it comtains something bootable.

These new findings may shed some additional and fresh light on everything
that has so far been discussed in this thread.

My sincere apologies to all participants for having asserted earlier that
absence of a USB stick from the Boot Priority list would absolutely insure
that the BIOS would not even try to boot off that device.  I genuinely
believed that, based on what I was told by Gigabyte tech suport, up until
just now.


Regards,
rfg


P.S.  If I am remembering all of the facts and data points correctly, it
now appears to me that (a) the presence or absence of the bootable flag
on _some_ partition on a given device is the thing which directly determines
whether or not the device in question will or will not show up in the
BIOS Boot Priority list, but that (b) presence or absense of a given device
on the BIOS Boot Priority list is *not* the primary determinant of whether
or not the BIOS will attempt to boot from the device in question, and
that rather, the BIOS will try to boot from a given mass storage device
if the class/category of the device is listed as one of the classes/
categories of devices that have been set as either the First or Second
or Third (categories of) Boot Devices _and_ if there are no higher priority
(First, Second, Third) categories of mass storage devices attached.

P.P.S.  Interestingly, since the BIOS apparently _will_ make an effort
to boot from USB mass storage devices (as long as "USB-HDD" is set as
one of the prioritized devices categories to boot from) even when the
device(s) in question DO NOT contain any partition which has the bootable
flag set for it (thus implying that the devices won't even show up in
the BIOS Boot Priority list) the question naturally arises:  How then
can you prioritize the boot order if you have two such USB sticks, i.e.
both _without_ any partition marked as "bootable".

The answer, with this BIOS, appears to be "You can't."  Rather,
some
additional tests that I performed just now indicate that in this case,
the BIOS will choose which of the multiple "no bootable partition" USB
mass storage devices it will actually try to boot from *first* strictly
based on the identity (i.e. number) of the specific USB port that each
device happens to be plugged into, i.e. there is apparently a fixed
hardware ordering of the USB _ports_, and that ordering determines boot
order for the multiple no-bootable-partition USB mass storage devices
in such cases.

On 01/17/2014 04:36 PM, Ronald F. Guilmette wrote:
> 
> To answer your question...
> 
>     *) with respect to Ultimate Boot CD (5.2.7), the answer is
"CHS".
> 
>     *) with respect to OpenELEC (3.2.3, I think), the answer is
"CHS".
> 
OK, it is "CHS"... which most likely means this is a geometry problem.
Bleh.  The one thing one can do with these is trying to find the "magic
geometry" that works on that machine, which may be 64/32 for example.
However, in CHS mode 1024x64x32 is only 1 GiB, and you might not be able
to boot from sticks that are bigger than that.
> 
> In any case, other experiments that I have just now done confirm that the
> real controlling factor(s), for this motherboard, are the settings of
> the BIOS's "First Boot Device", "Second Boot
Device", and "Third Boot
> Device" parameters.  If there is both a regular (sATA) drive and an
even
> potentially/theoretically bootable USB stick installed in the machine...
> even one that has only one partition, and where that partition might NOT
> have the bootable flag set... and if the First, Second, and Third Boot
> Devices are set as follows:
> 
> 	First: CDROM
> 	Second: USB-HDD
> 	Third: Hard Disk
> 
> then (in the absence of any CD/DVD) the BIOS will make an attempt to boot
> from the USB stick, before trying the sATA drive, EVEN IF the USB stick in
> question DOES NOT appear in the Boot Priority list.  And also, as one would
> expect, if the BIOS settings are instead:
> 
> 	First:	CDROM
> 	Second:	Hard disk
> 	Third:	USB-HDD
> 
> then (in the absence of any CD/DVD) then the BIOS will try to boot from the
> sATA Drive first, and will do so, assuming that it comtains something
bootable.
> 
> These new findings may shed some additional and fresh light on everything
> that has so far been discussed in this thread.
> 
Ah, it is an Award BIOS isn't it... I remember the Award BIOSes of that
era having that kind of issues.

	-hpa

In message <52D9E590.6090702 at zytor.com>, 
"H. Peter Anvin" <hpa at zytor.com> wrote:
>On 01/17/2014 04:36 PM, Ronald F. Guilmette wrote:
>> 
>> To answer your question...
>> 
>>     *) with respect to Ultimate Boot CD (5.2.7), the answer is
"CHS".
>> 
>>     *) with respect to OpenELEC (3.2.3, I think), the answer is
"CHS".
>> 
>
>OK, it is "CHS"... which most likely means this is a geometry
problem.
>Bleh.  The one thing one can do with these is trying to find the "magic
>geometry" that works on that machine, which may be 64/32 for example.
>However, in CHS mode 1024x64x32 is only 1 GiB, and you might not be able
>to boot from sticks that are bigger than that.
I don't pretend to understand any of this, so whatever you say I am
more than happy to accept without question.  (I do not have enough
knowledge to have any opinion of my own whatsoever on these matters.)
>Ah, it is an Award BIOS isn't it...
Yes.

Boot time message says "Award Modular BIOS..."


Regards,
rfg